Perhaps the most important analytical process of marketing is segmentation. By segmenting the market, one obtains a very clear understanding of customers; and segmentation ultimately provides a basis for clear and precise targeting and positioning.
But segmenting is also very difficult—and, especially without customer data, it is, I would admit, an art form in marketing.
You can become quite confused about segmentation if you read the popular press or most marketing textbooks, or look at countless Internet companies... because the term has been used to signify many things.
Typically, you'll find the term segmentation applied to demographics and lifestyles in consumer markets, and to size, industry and geography in business markets. On the Internet, people use age, gender, etc. for segmentation (or, worse yet, confuse segmentation with terms like one-to-one marketing—as though people are so unique that they have little in common).
It's all very confusing, but there is a way to make things clearer. The answer lies in the work of Russell Haley (Journal of Marketing, July 1968), who first used the term "benefit segmentation."
Also known as "Needs-Based segmentation," benefit segmentation is essentially the idea that customers should be segmented on the basis of their needs. Simply put, customers in different benefit segments have different needs.
Previously, we have explained benefits and tradeoffs, so you can use these terms to crystallize the idea of segments having different needs. The different segments allocate their 100 points differently across the various benefits. This allocation results in the needs of different segments clustering around different benefits.
Here is Haley's original segmentation of the toothpaste market. Notice how the segments seek a different cluster of benefits. Again, the segments trade off the possible benefits differently.
| The Sensory Segment | The Sociables | The Worriers | The Independents | |
| Principle Benefit Sought | Fruity or minty flavor, product appearance | White, bright teeth | Decay prevention, plaque and gum disease avoidance | Price |
| Demographic Correlate | Children | Young people | Families, older consumers | Men, large families |
| Behavioral Correlate | ? | Smokers | Heavy users of dental supplements | Heavy users of toothpaste |
| Personality Characteristics | High self-involvement | High sociability | Hypochondriacs | High autonomy |
| Lifestyle Characteristics | Hedonistic | Active | Conservative | Value-oriented |
You could easily imagine that the market for consumers buying on the Internet is similarly broken up into four or five segments with names like "Music Aficionados," "One-Stop Shoppers," "Techno-Media Types," etc. In fact, recent studies have shown that Internet users break into six categories, including "E-bivalent Newbies," "Clicks and Mortar," etc.
Two key ideas to note from this chart:
- First, the labels given to each segment are arbitrary. They are called the segmentation bases and are simply used to label a group of customers who care about a different cluster of benefits.
- Second, along the left column you will also see segment descriptors, or things that describe the customers in each segment. These descriptors may correlate with the basis of the segmentation (and if a descriptor is very highly correlated, it could be used as a basis for segmentation). In the above example, we can see that maybe lifestyle is highly correlated with the segmentation basis, so this could also serve as a good label for the segments.
Before getting into how one can think about the various ways to practically segment a market, let's first consider some key issues and questions:
Why should you segment by benefits, rather than the descriptors? There are three ways to answer this:
- It's the only way to have a clear message in the market.
- It's the only way to deliver what the customer wants.
- Marketing academics have not been successful at segmenting the markets differently and still finding meaningfully different segments.
Does this mean that descriptors are not used in marketing? No, they absolutely are used, but for a different purpose.
To see this, consider the following very simple example and how benefits are powerful and descriptors are useful. We have two fictitious segments in the cereal market and two demographics (young and old) and a product line (in colored bold) for each segment. Notice how clean this is in terms of a message (i.e., the benefits) for each segment. Note also how the descriptors are useful for the specific products names.
| Segment Names | Health-Conscious | Sweet-Tooth |
| Benefits | Healthy/Nutritious | Sweet/sugary |
| Young | Rough Riders | Honey Bears |
| Old | Bran Flakes | Sugar Blend |
Now, instead think about segmenting this market based on age (a typical way people might segment this market). What you would have is this:
| Segment Names | Old | Young |
| Benefits | ? | ? |
| Sugar Blend | Honey Bears | |
| Bran Flakes | Rough Riders |
But what message will you have here? That is, what are the benefits the old and young want? You wouldn't know!
Think this is not relevant to your business market? Here's another, real-world example.
Typically, firms that sell chips for cell phones (these are called DSPs) segment the market like this:
| Segment Names | Big Firms | Small Firms |
| Benefits | ? | ? |
But what benefits are big and small firms looking for? It turns out that this market can be segmented (at the level of the application) as follows:
| Segment Names | Innovators | Pragmatics | Quick & Easy |
| Benefits | Performance, support, upgradeability | Price/performance, applications support | Low price, turnkey solutions |
The point of these two examples is that by segmenting on the basis of age or size of firm, you would miss the fact, in the cereal example, that there is a segment of both young and old people who want sweet and sugary cereals. We should note that age segmentation is rampant on the Web; as a result, the messages that Internet companies have are muddled.
In the case of the DSP chips, segmenting by size of firm would miss the idea that within a given firm there might be both applications that are innovative (needing performance, support and upgradeability) AND "quick and easy" (needing low price and turnkey solutions). Plus, when segmenting on size of firm, you don't know what benefits to position your product on! That's the reason so many B2B companies have bland and ambiguous positions in the market.
So what are the bases for segmentation? Although it is difficult to determine what will segment a market into different segments based on different benefit tradeoffs, there are some useful ways of thinking about how this might be done. In any event, this is typically a reiterative process (i.e., trying one way, then another).
One way to start the process is to look at the various benefits and think about whether there are groups of customers who would care about different clusters of these benefits.
In my experience, this may lead you to the following bases for beginning to think about segmenting a market:
- Usage. Often, how customers use a product can result in their making tradeoffs across different benefits. For example, light users and heavy users of a product often care about different benefits.
- Application. Customers who apply a product in a mission-critical way often care about different benefits (for business buyers, this could be a way central to their business, but a cook may use a product that is central to a recipe as well).
- Prior experience with the product category. Often denoted as "expert" and "novice," these different types of prior experience usually highly correlate with different needs.
Often, in consumer markets, prior brand loyalty, buying situation (work versus entertainment) or sometimes lifestyle (as in the case of many cars) can be the basis of segmentation.
Without good data to help you, the best you can do is to begin trying to segment the market using some a priori idea (such as using usage, prior experience, etc.), then checking it to see whether the segments really care about different benefits. If not, try again, using different segmentation bases, and reiterate. You might look for combinations of bases as well.
You will know that you have a good segmentation if it meets the criterion that the customers in the different segments make tradeoffs differently. A good segmentation will also meet other criteria, such as these:
- The segments are measurable—that is, you can identify the size of the segment
- The segments are reachable—that is, you can reach the segment by media (often, this can be ascertained by looking at the segment descriptors)
Finally, some people get confused by thinking that segmentation means segment "targeting."
Segmentation is simply the analytical process of breaking the market into distinct segments. Targeting is a decision to go after a particular segment (see the tutorial), and this decision can only be made after you consider a number of other factors, including competitive response, customer perceptions, etc.
